# Efficient Evaluation-Time Uncertainty Estimation by Improved Distillation

arXiv: https://arxiv.org/abs/1906.05419 · HTML: https://mlsanity.com/paper/1906.05419/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Jun 2019.

## Built on by (most cited first, 2 of 2)

- [A Survey of Uncertainty in Deep Neural Networks](https://mlsanity.com/paper/2107.03342/) (2021)
- [Self-Knowledge Distillation via Dropout](https://mlsanity.com/paper/2208.05642/) (2022)

## Abstract (arXiv)

In this work we aim to obtain computationally-efficient uncertainty estimates with deep networks. For this, we propose a modified knowledge distillation procedure that achieves state-of-the-art uncertainty estimates both for in and out-of-distribution samples. Our contributions include a) demonstrating and adapting to distillation's regularization effect b) proposing a novel target teacher distribution c) a simple augmentation procedure to improve out-of-distribution uncertainty estimates d) shedding light on the distillation procedure through comprehensive set of experiments.